WO2003063017A2 - Carte topologique auto-adaptative a performance amelioree et a variation non monotone de la vitesse d'entrainement - Google Patents
Carte topologique auto-adaptative a performance amelioree et a variation non monotone de la vitesse d'entrainement Download PDFInfo
- Publication number
- WO2003063017A2 WO2003063017A2 PCT/IB2003/000170 IB0300170W WO03063017A2 WO 2003063017 A2 WO2003063017 A2 WO 2003063017A2 IB 0300170 W IB0300170 W IB 0300170W WO 03063017 A2 WO03063017 A2 WO 03063017A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- learning rate
- training
- function
- generated according
- weights based
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2137—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on criteria of topology preservation, e.g. multidimensional scaling or self-organising maps
Definitions
- the invention relates to Self-Organizing Feature Maps (SOFM) which are neural networks that transform an input of arbitrary dimension into a one or two dimensional discrete maps subject to a topological (neighborhood preserving) constraint and more particularly to such SOFMs in which the initial values of the weight vectors are random.
- SOFM Self-Organizing Feature Maps
- Neural networks occupy a large branch of research in machine intelligence.
- Artificial neural networks are information-processing devices inspired by the interconnected, parallel structure of animal brains. They take the form of software or hardware networks having collections of mathematical models that emulate some of the observed characteristics of nervous systems and analogies to adaptive biological learning. Generally, they are composed of large numbers of interconnected processing elements, which can be realized in software or hardware, that are analogous to the neurons of an animal brain. The connections between these processing elements are weighted in a fashion believed to be analogous to synapses.
- Training a neural network involves making adjustments to the "synaptic" connections that exist between the neurons (i.e., the values of the weights). Training is performed by exposure to a set of input/output data where the training algorithm iteratively adjusts the connection weights. These connection weights store the knowledge necessary to solve specific problems.
- Neural networks are being applied to greater and greater numbers of practical problems, including very complex ones. They are particularly suited to pattern recognition and classification problems having many inputs such as speech recognition, character and signal recognition, and functional prediction and system modeling where the physical processes are not understood or are highly complex.
- neural networks There are many types of neural networks. Some of the more popular include the multilayer perceptron, which is generally trained with the backpropagation of error algorithm, learning vector quantization, radial basis function, Hopfield, and SOFM. Some are classified as feedforward and others as recurrent (i.e., implement feedback) depending on how data is processed through the network. Another feature of neural networks is the mechanism by which it is trained. Some use a technique called supervised training while others are referred to as unsupervised or self-organizing. In supervised training, the network is guided by an instruction process while in unsupervised algorithms the data is clustered into similar groups based on the attributes that provide the inputs to the algorithms.
- the SOFM or Kohonen artificial neural network is a type of unsupervised neural network.
- unsupervised learning an untrained neural network is exposed to examples or input vectors and internal parameters adjusted.
- SOFMs all the neurons of the network receive the same input.
- the nodes engage in competition with their neighbors and at each stage of a self-learning process, the one with most activity "wins.” Learning is based on the concept of winner neurons.
- Unsupervised learning allows the objects to be grouped together on the basis of their perceived closeness in n- dimensional hyperspace (where n is the number of variables or observations made on each object). Such methods, then, although in some sense quantitative, are better seen as qualitative since their chief purpose is merely to distinguish objects or populations.
- SOFMs provide an objective way of classifying data through self-organizing networks of artificial neurons.
- Each node of the input layer may be connected (as indicated by connectors 120) to the entire set of nodes in the competition layer.
- each neuron may be connected to its eight nearest neighbors on a grid.
- the neurons store a set of weights (a weight vector) each of which corresponds to one of the inputs in the data.
- the objective of a Kohonen network is to map input vectors (patterns) of arbitrary dimension N onto a discrete map lying in a competition layer of arbitrary dimension, but typically of 1 or 2 dimensions.
- the algorithm adjusts weights so that patterns close to one another in the input space should be close to one another in the map: they should be topologically ordered.
- the learning process is as follows: First, the weights for each output unit are initialized, typically to random starting values. An iterative process is performed that ends when weight changes are negligible. For each of a number of input patterns a winning output node and all units in the neighborhood of the winner are identified and the weight vectors for all of these nodes updated. The winning output unit is simply the unit with the weight vector that has the smallest Euclidean distance to the input pattern.
- the neighborhood of a unit is defined as all units within some distance of that unit on the map (not in weight space). If the size of the neighborhood is 1 then all units no more than 1, either horizontally or vertically, from any unit fall within its neighborhood.
- the weights of every unit in the neighborhood of the winning unit are updated such that each unit in the neighborhood is moved closer to the input pattern. As the iterations are performed, the learning rate is reduced. If the parameters are well chosen the final network should capture the natural clusters in the input data.
- the factor that governs the size of the weight alterations is known as the learning rate.
- the above formula is one of a number that are known in the prior art and the invention to be discussed in subsequent section could make use of it or any other. As mentioned, the process of training continues until the changes in the weights fall below some predetermined value in successive iterations.
- the effect of the "learning rule” is to distribute the neurons evenly throughout the region of n-dimensional space populated by the training set.
- the neuron with the weight vector closest to a given input pattern will win for that pattern and for any other input patterns that it is closest to.
- Input patterns which allow the same node to win are then deemed to be in the same group, and when a map of their relationship is drawn a line encloses them.
- the resulting map it is possible to examine closely relationships between the items in the training set and visualize these relationships even for complex structures in high-dimensional input spaces.
- the rate of decrease may be exponential, linear, or some other and according to the prior art, the particular pattern has not generally been regarded as particularly important.
- the order phase the topological ordering of the weight vectors takes place.
- a long convergence phase follows and is associated with fine-tuning of the map.
- the learning rate is maintained at low values (well below 0.1, for example) and monotonically and progressively decreased with each iteration.
- the quality of a SOFM solution for any given problem is by no means assured. Where the input vectors are high-dimensional and complex, results may take a very long time to converge and may even produce poor end results.
- the learning rate is selected in a random or sporadic fashion. It has been found experimentally that this increases the rate of convergence in many instances. According to the prior art, the learning rate should be time-varying, but decreased monotonically. However, it has been discovered that, at least in some contexts, particularly in the case of higher dimensional input spaces, sporadic variation of the learning rate, at least during the initial stages, leads to improved performance.
- Fig. 1 is a diagrammatic representation of a self ordering feature map according to the prior art and consistent with embodiments of the invention.
- Fig. 2 is a flow chart representing an algorithm for implementing the invention according to an embodiment thereof.
- Fig. 3 is an illustration of selective random generation of learning rate parameters with progress of training of a SOFM.
- Fig. 4 is an illustration of another manner of selective random generation of learning rate parameters with progress of training of a SOFM.
- an algorithm for implementing the invention begins with the initialization of the weight vectors in step S10. Random values for all of the weight vectors are a typical choice.
- step S20 a sample input vector is drawn at random from a pool of input vectors being used for training the SOFM. The selected vector is applied to the input nodes and a winning competition layer node is identified in step S30 according to a minimum Euclidean distance value:
- Wi j is the/ ⁇ weight of the i th node
- I j is the/' 1 component of the input vector
- £>,- is the distance of the i th node.
- the node with the lowest distance value is the winner.
- a random value of the learning rate is generated. This learning rate is used to update the weights of the winning node and neighboring nodes.
- the winning node and all neighborhood nodes are then updated according to the following formula.
- ⁇ W ltJ a(I j - W u )sind /(2d)
- Wi j is the/ ⁇ weight of the i' h node
- a is the random learning rate
- I j is the/ ⁇ component of the input vector
- d is the distance between the current node and the winner.
- the generation of successive values of the learning rate can be performed in various ways consistent with the present invention.
- the values over which the random rates range should become smaller as the simulation progresses.
- the learning rate is a random value between bounds 161 and 162 which decrease gradually as the training process progresses. This needs not be a monotonic reduction in range, as illustrated at 160 in Fig. 3, but the learning rates preferably get smaller in value, and range over a smaller range, as the simulation progresses.
- the learning rate is varied in a similar range 160, but is varied cyclically or pseudorandomly.
- the learning rate varies around unity initially and falls to values near zero decreasing by several orders of magnitude.
- another alternative for generating may permit random variation (illustrated at 170) of the learning rate during the ordering phase of the training and switchover to monotonic reduction (illustrated at 180) of the learning rate for the convergence phase.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Machine Translation (AREA)
- Character Discrimination (AREA)
- Image Analysis (AREA)
- Measurement Of Resistance Or Impedance (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
- Feedback Control In General (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP03731793A EP1485814A2 (fr) | 2002-01-22 | 2003-01-21 | Carte topologique auto-adaptative a performance amelioree et a variation non monotone de la vitesse d'entrainement |
JP2003562812A JP2005516282A (ja) | 2002-01-22 | 2003-01-21 | 学習レートの非単調変化による性能の改善を伴う自己組織化特徴マップ |
KR10-2004-7011317A KR20040075101A (ko) | 2002-01-22 | 2003-01-21 | 학습 레이트의 비단조 변동에 의해 향상된 성능을 갖는자기조직화 형상지도 |
AU2003201125A AU2003201125A1 (en) | 2002-01-22 | 2003-01-21 | Self-organizing feature map with improved performance by non-monotonic variation of the learning rate |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/055,352 US6965885B2 (en) | 2002-01-22 | 2002-01-22 | Self-organizing feature map with improved performance by non-monotonic variation of the learning rate |
US10/055,352 | 2002-01-22 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2003063017A2 true WO2003063017A2 (fr) | 2003-07-31 |
WO2003063017A3 WO2003063017A3 (fr) | 2004-07-01 |
Family
ID=21997266
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/IB2003/000170 WO2003063017A2 (fr) | 2002-01-22 | 2003-01-21 | Carte topologique auto-adaptative a performance amelioree et a variation non monotone de la vitesse d'entrainement |
Country Status (8)
Country | Link |
---|---|
US (1) | US6965885B2 (fr) |
EP (1) | EP1485814A2 (fr) |
JP (1) | JP2005516282A (fr) |
KR (1) | KR20040075101A (fr) |
CN (1) | CN1623169A (fr) |
AU (1) | AU2003201125A1 (fr) |
TW (1) | TW200304096A (fr) |
WO (1) | WO2003063017A2 (fr) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2472440A1 (fr) | 2011-01-04 | 2012-07-04 | Siemens Aktiengesellschaft | Procédé et système pour le diagnostic du statut d'une installation |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4639784B2 (ja) * | 2004-12-06 | 2011-02-23 | ソニー株式会社 | 学習装置および学習方法、並びにプログラム |
JP4201012B2 (ja) * | 2006-04-06 | 2008-12-24 | ソニー株式会社 | データ処理装置、データ処理方法、およびプログラム |
JP2007280054A (ja) * | 2006-04-06 | 2007-10-25 | Sony Corp | 学習装置および学習方法、並びにプログラム |
US7814038B1 (en) | 2007-12-06 | 2010-10-12 | Dominic John Repici | Feedback-tolerant method and device producing weight-adjustment factors for pre-synaptic neurons in artificial neural networks |
CN103455843B (zh) * | 2013-08-16 | 2016-03-02 | 华中科技大学 | 一种反馈型人工神经网络训练方法及计算系统 |
US10769545B2 (en) | 2014-06-17 | 2020-09-08 | D-Wave Systems Inc. | Systems and methods employing new evolution schedules in an analog computer with applications to determining isomorphic graphs and post-processing solutions |
US10460237B2 (en) | 2015-11-30 | 2019-10-29 | International Business Machines Corporation | Neuron-centric local learning rate for artificial neural networks to increase performance, learning rate margin, and reduce power consumption |
US10599976B2 (en) * | 2016-11-07 | 2020-03-24 | International Business Machines Corporation | Update of attenuation coefficient for a model corresponding to time-series input data |
US11322073B2 (en) * | 2018-09-21 | 2022-05-03 | Dell Products, Lp | Method and apparatus for dynamically optimizing gamma correction for a high dynamic ratio image |
US11074344B2 (en) * | 2018-12-19 | 2021-07-27 | Intel Corporation | Methods and apparatus to detect side-channel attacks |
CN110705645A (zh) * | 2019-10-09 | 2020-01-17 | 重庆医药高等专科学校 | 一种基于sofm神经网络的英语教学质量评估方法 |
JP7420027B2 (ja) | 2020-09-10 | 2024-01-23 | 信越化学工業株式会社 | Euvマスクブランク用多層反射膜付き基板、その製造方法及びeuvマスクブランク |
JP2022045936A (ja) | 2020-09-10 | 2022-03-23 | 信越化学工業株式会社 | Euvマスクブランク用多層反射膜付き基板、その製造方法及びeuvマスクブランク |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2668625A1 (fr) * | 1990-10-30 | 1992-04-30 | Thomson Csf | Procede d'apprentissage automatique et reseau connexionniste multi-couches pour la mise en óoeuvre de ce procede. |
US5398302A (en) * | 1990-02-07 | 1995-03-14 | Thrift; Philip | Method and apparatus for adaptive learning in neural networks |
US5809490A (en) * | 1996-05-03 | 1998-09-15 | Aspen Technology Inc. | Apparatus and method for selecting a working data set for model development |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5235673A (en) * | 1991-04-18 | 1993-08-10 | International Business Machines Corporation | Enhanced neural network shell for application programs |
JPH05346915A (ja) * | 1992-01-30 | 1993-12-27 | Ricoh Co Ltd | 学習機械並びにニューラルネットワークおよびデータ分析装置並びにデータ分析方法 |
US6173275B1 (en) * | 1993-09-20 | 2001-01-09 | Hnc Software, Inc. | Representation and retrieval of images using context vectors derived from image information elements |
US5819245A (en) * | 1995-09-05 | 1998-10-06 | Motorola, Inc. | Method of organizing data into a graphically oriented format |
JP3647584B2 (ja) * | 1996-12-26 | 2005-05-11 | 富士通株式会社 | 学習型自己定位装置 |
US20010034686A1 (en) * | 1997-12-10 | 2001-10-25 | Eder Jeff Scott | Method of and system for defining and measuring the real options of a commercial enterprise |
US5933818A (en) * | 1997-06-02 | 1999-08-03 | Electronic Data Systems Corporation | Autonomous knowledge discovery system and method |
DE69826298T2 (de) * | 1998-12-29 | 2005-11-17 | International Business Machines Corp. | Verfahren und Vorrichtung zur Klassifikation von Netzwerkeinheiten in virtuellen LANs |
US6477469B2 (en) * | 2001-01-08 | 2002-11-05 | Halliburton Energy Services, Inc. | Coarse-to-fine self-organizing map for automatic electrofacies ordering |
-
2002
- 2002-01-22 US US10/055,352 patent/US6965885B2/en not_active Expired - Fee Related
-
2003
- 2003-01-21 JP JP2003562812A patent/JP2005516282A/ja not_active Withdrawn
- 2003-01-21 WO PCT/IB2003/000170 patent/WO2003063017A2/fr not_active Application Discontinuation
- 2003-01-21 AU AU2003201125A patent/AU2003201125A1/en not_active Abandoned
- 2003-01-21 CN CNA038025167A patent/CN1623169A/zh active Pending
- 2003-01-21 KR KR10-2004-7011317A patent/KR20040075101A/ko not_active Application Discontinuation
- 2003-01-21 EP EP03731793A patent/EP1485814A2/fr not_active Withdrawn
- 2003-01-22 TW TW092101351A patent/TW200304096A/zh unknown
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5398302A (en) * | 1990-02-07 | 1995-03-14 | Thrift; Philip | Method and apparatus for adaptive learning in neural networks |
FR2668625A1 (fr) * | 1990-10-30 | 1992-04-30 | Thomson Csf | Procede d'apprentissage automatique et reseau connexionniste multi-couches pour la mise en óoeuvre de ce procede. |
US5809490A (en) * | 1996-05-03 | 1998-09-15 | Aspen Technology Inc. | Apparatus and method for selecting a working data set for model development |
Non-Patent Citations (2)
Title |
---|
SCHNITMAN L ET AL: 'An efficient implementation of a learning method for Mamdani fuzzy models' PROCEEDINGS SIXTH BRAZILIAN SYMPOSIUM ON NEURAL NETWORKS, RIO DE JANEIRO, BRAZIL, 22-25 NOV 2000, LOS ALAMITOS, CA, USA, IEEE COMPUT. SOC, USA 2000, pages 38 - 43, XP010527545 ISBN: 0-7695-0856-1 * |
WEISHUI WAN ET AL: 'A new method to prune the neural network' PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, COMO, ITALY, 24-27 JULY 2000, LOS ALAMITOS, CA, USA, IEEE COMPUT. SOC, USA 2000, pages 449 - 454 VOL.6, XP010505028 ISBN: 0-7695-0619-4 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2472440A1 (fr) | 2011-01-04 | 2012-07-04 | Siemens Aktiengesellschaft | Procédé et système pour le diagnostic du statut d'une installation |
Also Published As
Publication number | Publication date |
---|---|
US6965885B2 (en) | 2005-11-15 |
JP2005516282A (ja) | 2005-06-02 |
EP1485814A2 (fr) | 2004-12-15 |
CN1623169A (zh) | 2005-06-01 |
TW200304096A (en) | 2003-09-16 |
WO2003063017A3 (fr) | 2004-07-01 |
AU2003201125A1 (en) | 2003-09-02 |
KR20040075101A (ko) | 2004-08-26 |
US20030140019A1 (en) | 2003-07-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sarkar et al. | A clustering algorithm using an evolutionary programming-based approach | |
Delashmit et al. | Recent developments in multilayer perceptron neural networks | |
Becker | Unsupervised learning procedures for neural networks | |
US6965885B2 (en) | Self-organizing feature map with improved performance by non-monotonic variation of the learning rate | |
US5050095A (en) | Neural network auto-associative memory with two rules for varying the weights | |
Kung et al. | A unified systolic architecture for artificial neural networks | |
US5832466A (en) | System and method for dynamic learning control in genetically enhanced back-propagation neural networks | |
US20220383126A1 (en) | Low-Rank Adaptation of Neural Network Models | |
Tiňo et al. | Learning and extracting initial mealy automata with a modular neural network model | |
CN110929798A (zh) | 基于结构优化稀疏卷积神经网络的图像分类方法及介质 | |
Bullinaria | Self organizing maps: fundamentals | |
Kothari et al. | Neural networks for pattern recognition | |
Jang et al. | Deep neural networks with a set of node-wise varying activation functions | |
Salehinejad et al. | Pruning of convolutional neural networks using ising energy model | |
Pal et al. | Fuzzy versions of Kohonen's net and MLP-based classification: performance evaluation for certain nonconvex decision regions | |
Wang et al. | Recurrent neural networks: Associative memory and optimization | |
Imada et al. | Application of an evolution strategy to the Hopfield model of associative memory | |
CN113205182B (zh) | 一种基于稀疏剪枝方法的实时电力负荷预测系统 | |
Inoue et al. | Efficiency of self-generating neural networks applied to pattern recognition | |
Martinini et al. | Structured Pruning in Deep Neural Networks with Trainable Probability Masks | |
Fu et al. | Evolving neural network ensembles using variable string genetic algorithm for Pattern Classification | |
Jang et al. | Deep Asymmetric Networks with a Set of Node-wise Variant Activation Functions | |
Douglas et al. | Relationships between Deep Learning and Linear Adaptive Systems | |
Shibata et al. | katsunarishibata@ gmail. com Oita University | |
Pearson et al. | Use of genetic and neural technologies in oil equipment computer-aided design |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SC SD SE SG SK SL TJ TM TN TR TT TZ UA UG UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2003731793 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2003562812 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 20038025167 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1020047011317 Country of ref document: KR |
|
WWP | Wipo information: published in national office |
Ref document number: 2003731793 Country of ref document: EP |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 2003731793 Country of ref document: EP |